- Home Archives for September 2011
Today's post at the Washington Post's Insider
looks at how the Redskins can take pressure off Rex Grossman. Plus:
- DeAngelo Hall does a lot of talking, but his play has been silent.
- Dallas' LB Sean Lee had a huge game, as I warned.
- One of the reasons the Skins are off to a good start is a very low penalty rate.
- Washington's run defense suddenly went from one of the best to one of the worst.
- Of the 10 FG attempts in Monday night's game, 4 were significant mistakes, particularly Dallas' kick on 4th and 1 from the WAS 9.
- Jason Campbell, the former Skins QB jettisoned in favor of Donovan McNabb and Grossman, has started off hot.
- Next week, WAS faces STL QB Sam Bradford, who has regressed in his second season. This doesn't speak well of STL's new OC, Josh McDaniels, the man who replaced Mike Shanahan in Denver.
The rankings and prediction model has been redone for 2011. Just like always, it's a team-efficiency logistic regression model. It's based on passing, running, turnover, and penalty efficiency. But now, running is represented by Success Rate (SR) rather than Yards Per Carry (YPC). SR correlates far better than YPC with winning games. YPC is too susceptible to a handful of relatively rare break-away runs. I think of running as a ‘jab’ and passing as a ‘cross’ or ‘uppercut’. The jab is a low-risk punch that doesn't expose your defenses, keeps your opponent off balance and guessing, and keeps him from purely defending against your cross. A good jab is a prerequisite, but the cross is what scores points and wins bouts. SR captures this aspect of the running game.
Running is more than that, of course. It’s essential in short yardage and inside the red zone, and when team has a lead, it burns clock and helps keep the ball out of their opponent’s hands in the 4th quarter. I believe the revised model better reflects the true inner workings of the sport.
Welcome to Week Three's edition of selected game recaps, where we take an objective look at the subjective best of the week's games, including:
- New England at Buffalo, in which the Bills make their presence as contenders official.
- Detroit at Minnesota, in which the Vikings prove allergic to leads again.
One of the biggest storylines of this early 2011 season is the obvious offensive explosion that has occurred, especially when it comes to quarterback play. Three QBs have already thrown for over 1000 yards, and nine quarterbacks are averaging over 300 yards per game through the air, and it doesn’t stop there. Of the 32 quarterbacks considered to be “qualified,” 23 of them have a standard QB Rating over 80. Eight signal callers currently sit above the 100 mark, and almost half of the leagues field generals come in over 90.
When QBs are universally playing out of their minds, how do we tell who’s really standing out and taking hold of the league? When league averages shift and context changes, we have to account for it in the way we evaluate players. If it’s easier to find a QB who can post a 90 rating then ever before, than those passers with ratings in the eighties aren’t so special any more. If you’re not convinced that we need to start comparing players to each other instead of just to the QB Rating metric, here is a chart that should change your mind.
4th and goals from the 15 might seem impossible, but they're not. The success rate is low but not zero. In fact it's about 14%. That's not a welcome prospect, but neither is handing the ball to an opponent who's up by 3 points with 4 minutes to play. And even if the conversion attempt fails, it's far from fatal. The calculus merely changes from needing a FG to tie, to needing a TD to win. Notice I said needing a FG to tie. Sure, TDs are harder to come by than FGs, but tying is not the object of the game. Put simply, TDs are more than half as likely as FGs in that situation, which makes the gamble for a tying FG a sucker's bet.
This is really meant for those with large monitors. Definitely not for your smart phone. It's the perfect companion for Red Zone channel junkies like myself. And it's best live.
Introducing The Wall.
Scott Chandler: The Most Important Football Player.
This week's edition of The Weekly League features
1. Mega Previews for the New England-Buffalo, Detroit-Minnesota, and Green Bay-Chicago football games.
2. An Interesting Fact™ regarding the Jay Cutler Sack Problem™.
3. Unflinching -- and mostly unwarranted -- devotion to Bills tight end Scott Chandler.
New England at Buffalo | Sunday, September 25 | 1:00pm ET
• In the event that you're curious about which receiver leads the 2-0 Bills in Expected Points Added (EPA), please be curious no longer -- because the answer is "tight end Scott Chandler."
• A couple other categories in which Scott Chandler leads the Buffalo receiving corps are EPA/P (1.12) and yards per target (8.8) -- i.e. pretty important things.
• All of which is to say: get Scott Chandler the ball, Buffalo Bills.
• A second note to the Buffalo Bills: try not to get Patriot quarterback Tom Brady the ball.
• Because he's the best at throwing it to his teammates, is why.
In this edition of selected game recaps, we take an objective look at the subjective best of week two, including:
- Oakland at Buffalo, in which Ryan Fitzpatrick dinks and dunks to victory.
- Philadelphia at Atlanta, in which Matt Ryan can't lose the game and Mike Kafka can't win it.
- Dallas at San Francisco, in which Tony Romo delivers an inspirational performance under adversity. No, really.
Today's post at the Washington Post's Insider looks at how rookie Ryan Kerrigan's impact has been both explosive and consistent. Plus, how poor is Rex Grossman's performance on 3rd downs?
Plus, finally, a new headshot photo!
I'd like to introduce the new writers/analysts here at Advanced NFL Stats. I was surprised by the response to the call for writers, which was both a blessing and a curse. It was great because there are so many talented people with an interest in football stats, but it also meant making some tough choices on who to include. I leaned strongly toward people with previous analytic and writing experience. So without further ado, here is the new cast:
One of the effects of the rule change in starting kickoff positions is that it makes onside kicks a considerably better bet. When you add a 15-yard penalty to the equation, teams should consider making an onside kick a fairly standard play.
There have already been four kicks from the 50 yard line this year. None have been onside. Onside kicks need to go at least 10 yards, but historically they have averaged 13 yards. A regular kick from the 50 yard line will almost certainly result in a touchback. According to the WP model, drives that start at the 20 yard line are worth 0.34 expected points. From the kicking team's perspective, that is -0.34 points. If the kicking team fails to recover the onside kick, they can expect their opponent to start the drive at the 37 yard line. That would be worth -1.25 EP. If an onside kick was successful, the kicking team would start at their opponent's 37 yard line. That would be worth 2.84 points. A successful onside kick is worth 3.18 points more than the value of the touchback (2.84 minus -0.34). An unsuccessful onside kick is worth -0.91 points (-1.25 minus -0.34). Clearly, a team should attempt an onside kick if they think they have anywhere near a 50% chance at success.
On 4th and 1 with 11:16 in the 4th qtr, 49ers kicker David Akers booted a 55-yard FG to put his team up by 10. Cowboys linebacker Keith Brooking was flagged for a 'leverage' penalty, meaning he used teammates to help elevate himself attempting to block the kick.
As I understand the rule, Harbaugh had the option of accepting the penalty for a 1st down at the Dallas 22 or deferring enforcement to the ensuing kickoff. He chose to forgo the 1st down, take the 3 points, and take the 15 yards on the kickoff.
Merely one of the jerseys Buffalo's Scott Chandler barely ever wore.
This week's edition of The Weekly League features
1. Previews for the Oakland-Buffalo, Green Bay-Carolina, and San Diego-New England games.
2. Some surprising facts about the Buffalo Bills and their Week One accomplishment.
3. Considerably more sass than frass.
Note: the author has replaced yards per run with success rate as the standard by which ORUN+ and DRUN+ are evaluated, on account of run success rate correlates with winning more strongly.
Oakland at Buffalo | Sunday, September 18 | 1:00pm ET
• Last week, the Buffalo Bills were the only team in the league to post above-average numbers in every one of the four factors (pass efficiency, pass efficiency allowed, run success rate, and defensive run success rate) en route to beating Kansas City, 41-7.
• Of course, one will note, a single game is a small sample.
• One will also note, however, that Kansas City was a better team than Buffalo last season and playing at home -- an advantage that would give one completely average team about a 57% chance of victory over another completely average team.
• Quarterback Ryan Fitzpatrick's relationship with tight end Scott Chandler proved particularly fruitful, as Chandler caught all five passes thrown to him, for an average of 12.6 yards per target (YPT) and 1.71 expected points added per play (EPA/P).
• This is the same Scott Chandler who had one reception in 14 career games entering the season.
Ok, not 4th down decisions exactly, but he hits the nail on the head here (via David Brooks):
People who have information about an individual case rarely feel the need to know the statistics of the class to which the case belongs.
Coaches want to know all about the particulars of certain situations: how fatigued the defense is, how the left guard has been playing today, how well the team has done in short yardage so far (sample size of 3 plays probably), how good the turf's traction is, and so on. But many coaches remain generally (and willfully) ignorant of the statistics of their larger situation.
Sure, all those little things matter, but not nearly as much as the 'class to which the case belongs', as Kahneman puts it.
Case in point.
One of the least well-known features on this site are the Match Up pages. They're a one-stop shop for comparing all the relevant stats between upcoming competitors. I created the pages originally as just a tool for myself, needing a quick and easy way to research the numbers for my posts at the NY Times and Washington Post.
There are tables to compare each team's conventional rate stats, offensive and defensive advanced stats, and offensive and defensive player stats. Unfortunately, it's quite an eye chart, so unless you're familiar with the stats and the tables, it might just seem like a wall of numbers.
You can get to them by clicking on the upcoming games in the scoreboard banner above, or you can always navigate to them directly at this address.
Today's post at the Washington Post's Redskins Insider looks at Grossman's performance and its context in his career. I also grade the o-line, look at a 4th and 5 conversion attempt, and show that a 14-point margin of victory can be misleading.
Damn, I need to get a new picture to those guys. That's just awful.
As a Baltimore guy, I fear the day the day when Ray hangs up his cleats. So I plotted his performance by year (since 2000 for regular season games) in terms of +WPA per game, +EPA per game, and Success Count (SC) per game, expecting to find some sort of noticeable decline. And that's exactly not what I found.
On all the individual defender stat pages here, I had included SR simply because I already had it coded for teams and offensive players. I thought it might be interesting and possibly reveal something, but no insight materialized. SR for defenders doesn't make sense for the same reason that total WPA or EPA, as derived from play-by-play descriptions, doesn't make sense.
The Redskins stunned the Giants on Sunday, winning 28-14 after shutting out Coughlin and crew in the second half. After a Rex Grossman fumble to start the 4th quarter, the Redskins blocked a 38-yard Lawrence Tynes field goal attempt and the momentum swung in their favor. Starting from their own 30, Washington went on a 6-minute, 10-play drive, culminating in a game-clinching touchdown pass to Jabar Gaffney.
Using a Markov model of a football drive, we can calculate the probability of a drive ending in any possible way: TD, field goal, turnover, punt, etc...
Here is a look at the Redskins final scoring drive:
Keith has agreed to be a contributor here at Advanced NFL Stats this season. He's been doing his own modeling and analysis at his site Drive-By-Football for several months. I was so impressed, I sought out Keith before my call for writers last week.
Keith is a true stat-head. Aside from Drive-By, he is the chief analyst at numberFire, a slick and smart fantasy-oriented site. He graduated Magna Cum Laude from Northwestern with a B.A. in math. Keith has worked with two NBA franchises - Oklahoma City Thunder and Philadelphia 76ers - doing statistical analysis and data management. In addition, he has worked with ESPN and the Wall Street Journal, contributing analysis across multiple platforms, and his sports analytics research has been featured at conferences across the nation.
Keith won't be confined to the GWP, WPA, AYPA environment. He has adapted his own Expected Points model (which has a clearer name--Net Expected Points).
So please welcome Keith. We can expect his first post very soon.
On another note, I have only begun reviewing all the responses for my call for writers. I was overwhelmed with all the interest. I intend to select an additional writer or two later this week.
The NFL play-by-play reports when players are injured on each play, or at least when the injury stops play so trainers can attend to the injured player. These are far from 100% all injuries suffered in the course of play, but they are the ones that tend to be significant or severe--ACLs, broken bones, separated shoulders, concussions--the kind of things that really worry players, teams, and the league.
With that information in hand, we can see the injury rates for each type of play, including kickoffs.
|Jay Cutler has fallen and doesn't want to get up.|
This week's edition of The Weekly League features
1. Heat-seeking previews for the Atlanta-Chicago, Minnesota-San Diego, and New England-Miami football games.
2. A total and purposeful omission of the Sunday night game between Dallas and the Jets.
3. A photo of Jay Cutler looking groggy -- i.e. basically the only type of photo of Jay Cutler on the internet.
Atlanta at Chicago | Sunday, September 11 | 1:00pm ET
• Against his better judgment, the author remains the sort of person who's pretty sure the Cutler-Martz Experiment will succeed.
• In fact, it sorta did succeed last year -- in that Cutler finished seventh among 31 qualified quarterbacks in yards per attempt, at 7.6 (compared to a league average of 7.0).
• Unfortunately, when considering net yards per attempt -- which also accounts for sacks -- Cutler finished 21st overall, at 6.0 (relative to the league average of 6.2).
• The Bears, of course, finished with a league-worst 10.7% sack rate.
• That Cutler's career sack rate is just 5.9% (i.e. about league average) and the Bears drafted lineman Gabe Carimi from Wisconsin bodes well for the team's passing attack (i.e. the thing that wins games).
If you run a website or blog, you can now include live win probability graphs on your own site. I've created a widget with some simple html code, and now it's available for all teams and all games.
Add the end of last season, I added another server that will handle a much greater traffic load. As regular visitors will attest, the demand for the live graphs would often crash the site, especially during nationally televised games. Hopefully, that's a thing of past. (The new link for live WP graphs is http://live.advancednflstats.com, but the old link will still work too.)
The WP widget is just like the smaller version of the graph I show on my main page during live games. Just add the following html snippet to your site.
Chose which team's game you want to show, and replace the bit that says ?team=GB with ?team=PIT or ...=BLT or ...=SEA or whichever team's abbreviation you like. That team's current or most recent game will automatically appear. This way, you won't have to update the code snippet each week. If you run a Jaguars fan site, just leave it at ?team=JAX all season.
I teased this last week, and now the switch has been flipped. All final Win Probability graphs now include individual player advanced stat box scores.
They're better seen than described, so go ahead and check out last year's Super Bowl.
For each team, I've included the top 2 QBs, top 3 RBs, top 5 WRs/TEs, and top 7 defenders in terms of WPA to be included. Advanced stats for the full team can be seen by clicking on the Team Offense or Team Defense table headers. Due to space limitations, I had to decide which stats to include, so I focused on the ones that you don't tend to see other places--things like targets for WRs.
So get to it:
Those interested should email me directly (see the About - Contact/FAQ menu link) with an introduction and links or attachments of 2 or more examples of your analysis and writing. If you have no previous experience, you can send ‘demo’ drafts of the kind of stuff you’d like to do.
In particular, I’m looking for contributors to ‘own’ a regular weekly assignment. For example (these are just random thoughts off the top of my head):
Win Probability Added (WPA) can't do everything, but one of the things it can do extremely well is tell us exactly how much of every win or loss was due to one component of a team. In this case, it can tell us how many wins the Peyton Manning passing game can account for. Although we can't really separate Manning from his blockers and receivers, we can nail down a hard number for the Colts passing game as a whole, of which Manning has been the central fixture.
Since 2000 (as far back as my data goes), Manning played in 176 regular-season games and accumulated a total of 43.0 WPA, for an average of 3.8 WPA per season. This equates to 0.24 WPA per game, which means that Manning (and his passing offense) would give an otherwise perfectly average team a 74% chance of winning a game. In other words, he would take an 8-win team and make them an 11.8-win team.